Study on human-machine interface based on mouth open/close motions

This paper describes a new human-machine interface based on mouth open/close motions. User attached the developed controller with the skin surface to observe the moth open/close states. The controller equipped a tiny magnet sensor by which the slight relative displacement between the upper and lower skin point of mouth could be measured. For a higher recognition ratio, the controller had a flexible structure to absorb the slight displacement caused not by operation but by mastication to eat something. Moreover, the misrecognition with speaking could be avoided by comparing the current mouth state with the previous one. Next, suggested method was applied for operation of the meal support equipment. The LED of four food trays automatically turned on and off in order at a constant period, and user could select the food on the tray by opening the mouth at a moment when the LED on target tray was illuminated. The manipulator scooped the selected food on the tray with the spoon and moved it to the front of the userpsilas mouth. As for the manipulation of scooping food, the humanpsilas manipulation could be divided into five processes, and the obtained humanpsilas characteristics was applied for the motion planning of the meal support manipulator.